Explicitly initializing a local variable.

Explicitly initializing a local variable.

In the previous commit, I made a wrong assumption that
policy::Schema::Normalize will always set the |changed|
parameter.  This has been fortunately caught by the Valgrind/FYI bot.

BUG=427513
Committed: https://crrev.com/1ffc24b7564d0363c65c3e9b6818c2360324da45
Cr-Commit-Position: refs/heads/master@{#317676}

[Łukasz Anforowicz, 2015-02-23 22:23:06]

lukasza@chromium.org changed reviewers:
+ sergeyu@chromium.org

[Łukasz Anforowicz, 2015-02-23 22:23:06]

Sergey, could you please take a look?

[Sergey Ulanov, 2015-02-23 22:26:35]

lgtm

[Sriram, 2015-02-23 22:32:19]

Can we actually have compiler warning for uninitialized locals converted to
build error? That way we can catch these quite easily in the future.

On Mon, Feb 23, 2015 at 2:26 PM, <sergeyu@chromium.org> wrote:

> lgtm
>
>
>
> https://codereview.chromium.org/947313003/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Łukasz Anforowicz, 2015-02-23 22:33:44]

The CQ bit was checked by lukasza@chromium.org

[commit-bot: I haz the power, 2015-02-23 22:39:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/947313003/1

[Sergey Ulanov, 2015-02-23 22:45:24]

On Mon, Feb 23, 2015 at 2:32 PM, Sriram Srinivasan <sriramsr@chromium.org>
wrote:

> Can we actually have compiler warning for uninitialized locals converted
> to build error? That way we can catch these quite easily in the future.
>

I believe useful warnings that compilers support should be already enabled
and clang will often warn for uninitialized variable. This particular bug
isn't something that can be detected by a compiler. Lucasz, did ASan also
catch this problem?

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[chromium-reviews, 2015-02-23 23:02:59]

No, ASan didn't catch the problem.

On Mon, Feb 23, 2015 at 2:45 PM, Sergey Ulanov <sergeyu@chromium.org> wrote:

> On Mon, Feb 23, 2015 at 2:32 PM, Sriram Srinivasan <sriramsr@chromium.org>
> wrote:
>
>> Can we actually have compiler warning for uninitialized locals converted
>> to build error? That way we can catch these quite easily in the future.
>>
>
> I believe useful warnings that compilers support should be already enabled
> and clang will often warn for uninitialized variable. This particular bug
> isn't something that can be detected by a compiler. Lucasz, did ASan also
> catch this problem?
>
>


-- 
Thanks,

Lukasz

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[commit-bot: I haz the power, 2015-02-23 23:28:13]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-02-23 23:29:10]

Patchset 1 (id:??) landed as
https://crrev.com/1ffc24b7564d0363c65c3e9b6818c2360324da45
Cr-Commit-Position: refs/heads/master@{#317676}

[Łukasz Anforowicz, 2015-02-23 23:44:34]

Actually I was wrong - the issue was eventually caught by an MSan bot:
http://build.chromium.org/p/chromium.memory.fyi/builders/Linux%20MSan%20Tests...

[Łukasz Anforowicz, 2015-02-24 18:22:40]

A quick status update / follow-up to earlier questions:
1. We already run with -Wall
(https://code.google.com/p/chromium/codesearch#chromium/src/build/common.gypi&...)
2. I verified that -Wall catches simple case of unininitalized variables (based
on the one at
http://stackoverflow.com/questions/22518944/clang-not-reporting-uninitalized-...)
3. I verified that -Wall -Wextra doesn't catch my specific case.
4. I added more data to an already existing feature request for warning in this
specific case: http://llvm.org/bugs/show_bug.cgi?id=17558